Until the end of the last century, influenza was considered a serious illness only in the elderly and people of any age with certain chronic medical conditions leading to an increased risk of complications.^1–3 For this reason, influenza vaccination was recommended only in subjects more than 64 years of age and in children as well as adults with high risk medical conditions.^4,5

However, it has recently been demonstrated that, during influenza periods, healthy children less than 2 years of age had a significantly greater risk of hospitalisation for acute respiratory disease than older healthy children and a risk similar to that of children 5–15 years of age with high risk medical conditions.⁶ Moreover, other studies have shown excess numbers of outpatient visits and antibiotic prescriptions among healthy day care and school aged children.^7–9 Finally, several reports have shown that children are the major route of transmission of influenza viruses to household contacts, thus highlighting the potential of influenza to affect the quality of life of children and their families.^10–13

Recent findings on the impact of influenza in healthy children have led the US Advisory Committee on Immunization Practices to encourage influenza vaccination in all subjects aged 6–23 months.¹⁴ However, the interpretation of many studies is confounded by the lack of laboratory confirmation of influenza-like illnesses, the co-circulation of respiratory syncytial viruses, and the fact that they are retrospective.^6–8 Furthermore, the impact of influenza may vary from year to year, and be influenced by circulating strains and protective antibody levels in the population.¹⁴ Consequently, further virological and epidemiological data evaluating the global health of children and economic burden of influenza in healthy children are needed to define the best strategies for influenza prevention in paediatric subjects. This prospective study was designed to evaluate the burden of laboratory confirmed influenza in healthy children and their households.

METHODS

Study design

This was a prospective, multicentre study of children with respiratory tract infections in Italy. The patients were enrolled in four emergency departments (EDs) and by five primary care paediatricians (PCPs) in different Italian municipalities between November 1, 2001 and April 30, 2002. The EDs were located in Milan, Genoa, Florence, and Naples, and the PCP clinics in Milan and Genoa. Most of the children enrolled in EDs were taken to the hospital directly by their parents and were not referred by PCPs. The study protocol was approved by the Institutional Review Board of each centre.

Study population

During the study period, we enrolled all subjects younger than 14 years of age without chronic medical conditions presenting on 2 days of the week (Wednesday and Sunday in EDs or Tuesday and Thursday in PCP clinics) with signs and/or symptoms of respiratory tract infection, diagnosed as the presence of at least one of the following symptoms: runny nose, nasal congestion, sore throat, cough, earache, wheezing, and/or shortness of breath, regardless of fever (defined as an axillary temperature ⩾38°C).¹⁵ The exclusion criteria included concomitant chronic diseases at increased risk for influenza complications.^1–3 Only one child per family was included in the study. Written informed consent of a parent or legal guardian was required, and the older children were asked for their assent. There were no refusals to participate.

Enrolment and evaluation of patients

Upon enrolment, systematic recordings were made of the patients’ demographic characteristics and medical history using standardised written questionnaires. The questions included: detailed signs and symptoms of the acute episode of respiratory tract infection; laboratory and/or radiological examinations; prescribed drug therapy; previous administration of influenza vaccine; family size and number of siblings; parents’ education and occupation; family living conditions; and information about child care attendance. After a complete physical examination, the children were classified into six disease groups based on well established criteria¹⁶: common cold, pharyngitis, acute otitis media, croup, acute bronchitis, and pneumonia. When signs and symptoms of more than one disease were present, children were classified in the more severe disease group. The diagnosis of acute otitis media was performed with pneumatic otoscopy and the diagnosis of pneumonia with chest radiograph.

Nasopharyngeal samples were collected at enrolment by Virocult swab (Medical Wire and Equipment, Corsham, UK) to obtain specimens for influenza virus isolation and RNA detection. The samples were immediately sent by courier to a central laboratory (Department of Health Sciences, University of Genoa, Genoa), where they were divided into two fractions: one was inoculated into MDCK cells for isolation and the other was used for detection, typing, and subtyping by gene amplification as previously described.^17–20 The isolated strains were further identified by a haemoagglutination inhibition test using ferret post-infection sera at the WHO Influenza Centre, London, UK. Viral RNA was extracted using QIA techniques according to the manufacturer’s instructions (RNeasy Minikit, Qiagen, Valencia, CA). Reverse transcription (RT) and polymerase chain reaction (PCR) were performed using standard methodologies.^18–20

The medical history of the children was re-evaluated 5–7 days after enrolment and until the resolution of their illness by means of interviews and clinical examination by trained investigators using standardised questionnaires. During this evaluation, information regarding respiratory illnesses and related morbidity among households was also obtained. Parents or legal guardians were asked to answer a list of questions regarding the outcome of the disease in their children (for example, final diagnosis, administered medication, hospitalisation, duration of signs/symptoms of the illness, medical visits, examinations, number of school days lost) and the involvement of other family members (for example, respiratory diseases in household contacts, medication, hospitalisation, medical visits, number of work days lost by parents to care for their ill children and their own respiratory diseases, number of days of domestic help required to care for ill children). All data of the study children and their households were verified from medical records.

Statistical analysis

The data of all patients were analysed using SAS for Windows v. 12 (SAS Institute, Cary, NC, USA), and comparisons made between influenza positive and influenza negative cases, influenza positive children enrolled in EDs and by PCPs, influenza A and influenza B cases, and influenza positive children by age (<2 years, 2–5 years, >5 years). A p value of <0.05 was considered statistically significant for all tests. Parametric data were compared using analysis of variance (ANOVA); abnormally distributed or non-parametric data were analysed using the Kruskal-Wallis test. Categorical data were analysed using contingency analysis and the χ² or Fisher’s test.

RESULTS

Epidemiology

The study included 3771 children younger than 14 years of age (1990 males; median age: 2.5 years), 2970 (78.8%) enrolled in EDs, and 801 (21.2%) by PCPs. Influenza virus was demonstrated in 352 cases (9.3%): 260/2970 (8.7%) children enrolled in EDs and 92/801 (11.5%) enrolled by PCPs (p = 0.022). Influenza A viruses were identified in 183 cases (52.0%: 129 H3N2, 70.5%; 54 H1N1, 29.5%) and influenza B virus in 169 (48.0%). In all cases, influenza viruses were detected by PCR and in 319/352 (90.6%) samples isolation of influenza viruses was also obtained. No significant difference was observed in the epidemiology of influenza infection in the different municipalities.

Figure 1 shows the distribution of influenza positive swabs. Considering all the influenza viruses together, the peak incidence of positive cases (at least 10% of the total number of positive influenza virus tests) was between the 3rd and 10th weeks of 2002, the highest incidence occurring in the 7th week (17.4% in EDs and 35.3% in PCP clinics) (p = 0.019). Influenza B virus circulated earlier than influenza A viruses, peaking between the 3rd and 7th weeks of 2002 versus the 8th and 10th weeks.

Characteristics of the study patients

Table 1 shows the general characteristics of the influenza positive and influenza negative children; the only significant difference was median age, which was higher in the influenza positive children. However, as there were 109, 167, and 76 cases of influenza A and B in children <2 years, 2–5 years, and >5 years of age, respectively, 78.4% of all influenza cases were in children younger than 5 years of age.

The influenza positive children seen by PCPs had a significantly older mean (SD) age than those enrolled in EDs (5.48 (3.38) v 3.43 (2.56) years, p<0.0001), and were significantly more frequently affected by influenza B infections (52/801, 6.5% v 117/2970, 3.9%, p = 0.002). Influenza A cases were equally distributed in the different age groups (<2 years, 2–5 years, >5 years), whereas the influenza B cases had a significantly older mean age than the influenza A cases (4.81 (3.24) v 3.19 (2.38) years, p<0.0001) (fig 2). There were no other significant differences in the general characteristics of the children enrolled in EDs or by PCPs, or between influenza A and influenza B positive subjects.

Table 2 shows the data on clinical presentation, the diagnostic methods used at enrolment, therapeutic approaches, and clinical outcomes.

The clinical presentation of respiratory tract infection was similar among the influenza positive and influenza negative children, but fever and croup were significantly more frequent among the former, and acute otitis media was significantly more frequent among the latter. Croup was significantly more frequent in the influenza positive children enrolled by PCPs than in those enrolled in EDs (36/92, 39.1% v 16/260, 6.1%, p<0.0001), but there was no other significant between group difference. The children with influenza B virus had croup significantly more frequently and pharyngitis significantly less frequently than those with influenza A viruses (35/169, 20.7% v 17/183, 9.2%, p = 0.004, and 69/169, 40.8% v 99/183, 54.0%, p = 0.017); no other significant differences were detected. The influenza positive patients aged <2 years had fever at any time during influenza illness significantly less frequently than those aged 2–5 years (88/109, 80.7% v 155/167, 92.8%, p = 0.004) or >5 years (88/109, 80.7% v 72/76, 94.7%, p = 0.011); no other age related differences were observed.

The diagnostic methods used at enrolment were similar in influenza positive and influenza negative children, influenza positive children enrolled in EDs or by PCPs, influenza A and B cases, and influenza positive children of different ages.

In terms of therapy, the influenza positive children received antipyretics significantly more often and for longer periods than their negative counterparts, whereas there was no significant difference in antibiotic prescription and use. Despite the international campaign for control of injudicious antibiotic use, antibiotic prescriptions were dispensed for more than 50% of the children in both the groups. Antiviral agents were not used in any case. No difference in medication usage was observed between the influenza positive children enrolled in EDs or by PCPs, or between influenza A and influenza B cases. The influenza positive children aged <2 years were prescribed antipyretics significantly less frequently than the children aged 2–5 years (73/109, 66.9% v 133/167, 80.8%, p = 0.026) or >5 years (63/76, 80.3%, p = 0.024); no other age related differences were observed.

The influenza positive children had a longer duration of fever and school absence (evaluated among children with full time child care attendance) than the influenza negative children, whereas the prevalence of hospitalisation, its duration, and the number of additional medical visits were similar. No significant differences were observed between the influenza positive children enrolled in EDs or by PCPs, between influenza A and influenza B cases, or between the different age groups.

Socioeconomic effect of influenza among households

Table 3 shows that the households of the influenza positive children had significantly more respiratory tract infections than those of the influenza negative children, received significantly more antipyretics and antibiotics, needed significantly more medical visits, missed significantly more work or school days, and needed significantly more help at home to care for ill children.

No significant differences in socioeconomic impact were observed between the households of influenza positive children enrolled in EDs or by PCPs. Home help to care for ill children was required for a significantly longer time in influenza B than in influenza A cases (1.30 (2.35) days v 0.77 (1.82) days, p = 0.0016). No other significant between group difference was found.

DISCUSSION

To the best of our knowledge, this is the first prospective study simultaneously evaluating the effect of laboratory confirmed influenza on healthy children and their household contacts, comparing influenza infection in different paediatric practice settings, analysing differences between influenza A and influenza B in a large group of children, and studying the age related characteristics of influenza infection. The main findings seem to be those indicating the substantial impact of influenza on healthy children and their families. Like previous studies,^21–24 we found that healthy children attending day care and school are the most frequently affected by influenza and their illness is often associated with a high disease burden in families.

The fact that influenza in healthy children can cause relevant social problems is shown by the observation that our influenza positive children missed significantly more school days than those with respiratory disease due to different pathogens. These data support the wider paediatric use of influenza vaccination and suggest that the current recommendation from the United States encouraging the use of influenza vaccination in healthy children aged <2 years¹⁴ should be extended to older children.

The socioeconomic impact of influenza in healthy children on household contacts, which seems to be significantly greater than that due to other respiratory infections, is demonstrated by the fact that the parents and siblings of our influenza positive children fell ill significantly more frequently than those living with children suffering from respiratory diseases of different etiology, received more antibiotic and antipyretic prescriptions, and required more medical visits. They were also more frequently absent from work or school and required home help for a longer time to care for ill children. These findings extend some previously published data,^15,25,26 and indicate that influenza in healthy children considerably affects their families, thus suggesting that influenza prevention may have a significant household impact.

The importance of influenza in paediatric practice is highlighted by the number of peak period visits to EDs and PCP clinics for respiratory diseases due to influenza viruses, with only marginal differences between the two settings. Although in our study children were enrolled during only 2 days per week, the long duration of the study period strengthens our conclusions. Moreover, the considerable influenza related medical burden even in otherwise healthy children is confirmed by the high rate of hospitalisation among our laboratory confirmed cases, regardless of age. Although no other viruses were detected, it is probable that a significant number of the influenza negative cases were due to respiratory syncytial virus (whose epidemiology in Italy is similar to that of influenza viruses), which is the most frequent cause of hospitalisations in the first years of life.^27–29 Furthermore, most of our hospitalised influenza positive children suffered from lobar pneumonia, thus underlining the possible role of influenza viruses in predisposing healthy subjects to serious diseases and possible bacterial superinfections.^30,31

Our data also add further information concerning the clinical picture of paediatric influenza. We confirm that fever is a major component of influenza presentation,^13,32,33 as it was more frequent in our influenza positive children than in those with respiratory diseases due to other agents. However, our findings also show that this sign may be absent in children aged <2 years, who are at higher risk for influenza related complications. This suggests that, during epidemic periods, respiratory diseases in younger children should be carefully monitored even when presenting as mild illness. We also confirm that croup is associated with influenza infection,^33,34 especially when influenza B virus is involved. Unexpectedly,^9,35–37 we found a lower incidence of acute otitis media in influenza positive than influenza negative children. This may be related to a high incidence of respiratory syncytial virus (RSV) infection among influenza positive children, as RSV is the virus that is most frequently associated with otitis media, or might be because the frequency of acute otitis media is related to age,^38,39 and the mean age of our influenza negative children was less than that of our influenza positive patients, with a significant number being aged <3 years; on the other hand, no other characteristic appeared different between the two groups.

In conclusion, our results demonstrate that influenza in otherwise healthy children has a significant effect on the children themselves and their household contacts. The prevention of influenza among these children may significantly reduce the number of sick subjects, the prevalence of hospitalisation, and the circulation of infection within households. Although any change in the policy of influenza vaccination requires a balanced evaluation of many aspects, our results suggest that prevention strategies should also focus on healthy children regardless of age because of their role in disease transmission.